2016-2017 GRaNT Computer Rankings

[D2D]

Re: 2016-2017 GRaNT Computer Rankings

Rounding off:
Wisconsin 4, Robert Morris 1 (I don't think it will be this close)
Boston College 2, St. Lawrence 2 (to be decided in overtime?)
Clarkson 3, Cornell 2 (not taking into account the Big Red's 2nd leading scorer is out)
Minnesota-Duluth 3, Minnesota 2 (OK, 2.5 to 2.3 - we've got a shot!)

[TonyTheTiger20]

Re: 2016-2017 GRaNT Computer Rankings

Boston College 2, St. Lawrence 2 (to be decided in overtime?)

Oh God nooooooooooooooooooooooooooooooooooooooooooooooooo oooooooooooooooooo

[TonyTheTiger20]

Re: 2016-2017 GRaNT Computer Rankings

Oh, one other thing -- we can debate whether a 4-3 win is better than a 6-4 win or whatever you want, but the theory behind using this methodology is to reward teams a team that wins 4-0 or 5-1 or some other big win, and determine that team to be better than another team who might have the same win/loss result, but might win a 2-1 or 3-2 close one.

I'm not saying this is better than KRACH or CHODR or any other methodology out there (except for PWR and RPI, those are stupid), just that it's another methodology to go alongside them, depending on what level of the results (i.e. wins/losses, goals, or even shots) you want to analyze. I'm still the world's biggest KRACH addict -- which is why I'm so excited about my system, because the theory behind the calculation is very similar, it's just using a different level of game results.

[TonyTheTiger20]

Re: 2016-2017 GRaNT Computer Rankings

Bump

[ARM]

Re: 2016-2017 GRaNT Computer Rankings

Bump

I did send an email requesting Spam removal. In the meantime, thanks for moving actual topics to the front.

[TonyTheTiger20]

Re: 2016-2017 GRaNT Computer Rankings

I did send an email requesting Spam removal. In the meantime, thanks for moving actual topics to the front.

Cheers! Noticed someone did it in the men's forum and figured it would take less time to do it in here

[vicb]

Re: 2016-2017 GRaNT Computer Rankings

Oh, one other thing -- we can debate whether a 4-3 win is better than a 6-4 win or whatever you want, but the theory behind using this methodology is to reward teams a team that wins 4-0 or 5-1 or some other big win, and determine that team to be better than another team who might have the same win/loss result, but might win a 2-1 or 3-2 close one.

I'm not saying this is better than KRACH or CHODR or any other methodology out there (except for PWR and RPI, those are stupid), just that it's another methodology to go alongside them, depending on what level of the results (i.e. wins/losses, goals, or even shots) you want to analyze. I'm still the world's biggest KRACH addict -- which is why I'm so excited about my system, because the theory behind the calculation is very similar, it's just using a different level of game results.

I think all the work you have done is amazing. One other thought though, right now there is no incentive for teams to try and score as many goals as possible in order to get as high a ranking as possible. I would be afraid that implementing this type of scoring system would lead to teams “running up” the score. Just sayin.

And I am having all kinds of issues typing responses on the USCHO threads. I type a letter and then the next letter I want to type I have to position my mouse to the right of the previous letter, left click the mouse button so now I type the next letter and so on and so on.
To type this response I had to do it in Word and then copy this to the USCHO thread.

[TonyTheTiger20]

Re: 2016-2017 GRaNT Computer Rankings

I think all the work you have done is amazing. One other thought though, right now there is no incentive for teams to try and score as many goals as possible in order to get as high a ranking as possible. I would be afraid that implementing this type of scoring system would lead to teams “running up” the score. Just sayin.

No you're definitely right, I agree. I don't think any rating system should be actually implemented for a purpose (i.e. to select the tournament) if it involves margin of victory because it does incentivize blowouts. But, tools that do (such as some of the versions of Jeff Sagarin's ratings, and GRaNT) can definitely be useful as an unofficial means of trying to determine which teams are better than others.

[ne7minder]

Re: 2016-2017 GRaNT Computer Rankings

I did send an email requesting Spam removal. In the meantime, thanks for moving actual topics to the front.

I have sent messages 2 or 3 times so it is nice to know others have too. Now if it just appeared as if anyone at USCHO gave a rip . . .